Abstract
This letter discusses mixed-mode magneto tunneling junction (m-MTJ)-based restricted Boltzmann machine (RBM). RBMs are unsupervised learning models, suitable for extracting features from high-dimensional data. The m-MTJ is actuated by the simultaneous actions of voltage-controlled magnetic anisotropy and voltage-controlled spin-transfer torque, where the switching of the free-layer is probabilistic and can be controlled by the two. Using m-MTJ-based activation functions, we present a novel low area/power RBM. We discuss online learning of the presented implementation to negate process variability. For MNIST hand-written dataset, the design achieves ~96% accuracy under expected variability in various components.
Highlights
Restricted Boltzmann machine (RBM) is a two-layered stochastic recurrent neural network [1] capable of unsupervised learning and feature extraction
Unlike a typical neural network, the activation functions in RBM are stochastic and produce an output ‘0’ or ‘1’ with a probability determined by the network weights and the applied pattern
This work discusses the potential of mixed-mode magneto-tunneling junctions (m-MTJ) to simplify the activation functions in RBM. m-MTJ is a two terminal magnetic device where the probability of switching of
Summary
Restricted Boltzmann machine (RBM) is a two-layered stochastic recurrent neural network [1] capable of unsupervised learning and feature extraction. The two layers of the RBM are visible layer to which the test/training sets are applied and the hidden layer which is the feature extractor. Each node of RBM has a probability to be in state 1 according to (1b) and (1c) where σ is the logistic sigmoid function. Unlike a typical neural network, the activation functions in RBM are stochastic and produce an output ‘0’ or ‘1’ with a probability determined by the network weights and the applied pattern. A physical implementation of stochastic activation function in RBM is challenging. This work discusses the potential of mixed-mode magneto-tunneling junctions (m-MTJ) to simplify the activation functions in RBM.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
